Thread: Re: error-free disabling of individual child partition

Re: error-free disabling of individual child partition

From
"Zeugswetter Andreas DCP SD"
Date:
> Added to TODO:
>
>         o Add ALTER TABLE tab ADD/DROP INHERITS parent

Sounds good, more inline with add/drop constraint.

>           pg_attribute.attislocal has to be set to 'false' for ADD,
and
>> attislocal: If you set this to False, you wouldn't be able to set it
>> back again.

Just curious, why is that ?

>           pg_attribute.attinhcount adjusted appropriately

Do we have a plan on what to do with RI constraints on the parent
when doing DROP INHERTITS ?
Seems all FK tables would need to be checked for keys in the local
table.
(Not sure whether we have PK on tab*, but it seems you could do it when
the
partitioning column is part of the PK)

Andreas


Re: error-free disabling of individual child partition

From
Hannu Krosing
Date:
Ühel kenal päeval, K, 2006-05-24 kell 09:56, kirjutas Zeugswetter
Andreas DCP SD:
> > Added to TODO:
> > 
> >         o Add ALTER TABLE tab ADD/DROP INHERITS parent
> 
> Sounds good, more inline with add/drop constraint.
> 
> >           pg_attribute.attislocal has to be set to 'false' for ADD,
> and
> >> attislocal: If you set this to False, you wouldn't be able to set it
> >> back again.
> 
> Just curious, why is that ?
> 
> >           pg_attribute.attinhcount adjusted appropriately
> 
> Do we have a plan on what to do with RI constraints on the parent
> when doing DROP INHERTITS ? 

I think this is out of scope of this discussion, as we do not currently
support RI over inherited tables. IOW RI is always on one table ONLY.

-- 
----------------
Hannu Krosing
Database Architect
Skype Technologies OÜ
Akadeemia tee 21 F, Tallinn, 12618, Estonia

Skype me:  callto:hkrosing
Get Skype for free:  http://www.skype.com